percolationindicator: Capturing the dynamics of a network in a single parameter:...

Description Usage Arguments Value Note Author(s) References

Description

This method captures the dynamics of a network in a single parameter: the percolation indicator. The indicator is computed by combining intensive longitudinal binary data of a system with its network structure (with temporal relationships). The dynamics of the network is modeled by two independent Poisson processes. The ratio of the parameters of these processes (infection rate and recovery rate) is used to obtain an indication of the development of the network over time. When the percolation indicator pi <= 1, activity in the network will die out. When pi > 1, the network will remain infected indefinitely. Can deal with binary data and an unweighted adjacency matrix.

Usage

1
PercolationIndicator(data, adjacency, separate = FALSE, ...)

Arguments

data

A matrix with multiple observations of one system. The dimension of the matrix is nobs x nvars; each row is a vector of observations of the variables on a specific time point. Must be longitudinal binary data.

adjacency

An unweighted adjacency matrix. The dimension of the matrix is nvars x nvars. The adjacency matrix may also be (2 x nvars) x (2 x nvars), containing temporal relationships.

separate

Logical. Can be TRUE or FALSE. The default is FALSE, resulting in one percolation indicator that applies to the whole network. When TRUE, percolation indicators per node of the network are returned.

...

Currently not used.

Value

PercolationIndicator returns a "PercolationIndicator" object that contains the following items:

inf.rate

The infection rate (of the network or per node of the network).

recov.rate

The recovery rate.

perc.ind

The percolation indicator (of the network or per node of the network).

Note

See also my website: http://cvborkulo.com

Author(s)

Claudia D. van Borkulo

Maintainer: Claudia D. van Borkulo <[email protected]>

References

van Borkulo, C. D., Kamphuis, J. H., and Waldorp, L. J. (manuscript submitted for publication). Predicting behaviour of networks: the contact process as a model for dynamics.

Fiocco, M. and van Zwet, W. R. (2004). Maximum likelihood estimation for the contact process. Lecture Notes-Monograph Series, 309 - 318.

Grimmet, G. (2010). Probability on graphs: random processes on graphs and lattices, 1. Cambridge University Press.

Harris, T. E. (1974). Contact interactions on a lattice. The Annals of Probability, 2, 969 - 988.


cvborkulo/PercolationIndicator documentation built on May 12, 2017, 2:23 p.m.